12 



are at first wliite, but very soon after the formation of the eonidia beds 

 they become colored, the variable tint raDging from pink, pale tlesh 

 color or pale salmon to deep salmon. When germinated in water or 

 under acid or alkaline agar or iu very moist air, producing eonidia 

 indistingaishable from those borne by the internal fungus (PI. II, 10; 

 PL 111,11 in part).' Immature spores from the eonidia beds are, of 

 course, much shorter than the measurements given above, and for a 

 time are non-septate or only 1-septate. When accidentally knocked 

 off, such spores appear to be still capable of growth and septation. An 

 effort was made to identify this external conidium with some one of the 

 many described forms of Fusarium, but without success. The above 

 measurements are derived from hundreds of spores and are believed to 

 give nearly the limits of variability, but it would be quite possible to 

 make half a dozen species of Fusaria from the material I have had in 

 hand, if only a few measurements were made and a hasty description 

 written out, as is frequently done. Indeed, there is no doubt that the 

 literature of systematic mycology contains much of such rubbish. One 

 has only to study critically a few members of this group and then turn 

 to the descriptions iu Saccardo's Sylloge Fungorum to be convinced 

 of it. Not only is the identification of Fusaria from descriptions usu- 

 ally impossible, but very often with the specimens themselves the case 

 is little better, owing to the fact that many Hypocreaceous fungi bear 

 eonidia which have been put .into the form-genus Fusarium and which 

 so closely resemble each other (see the plates in Vol. Ill of Tulasne's 

 Carpologia, and iu Heft 10 of Brefeld's Untersuchungen) that mor- 

 phology alone is of very little assistance in their identification. In 

 many cases, at least, the only safe way is to make cultures and inocu- 

 lations if the fungus is a well-marked parasite ; or to obtain the less 



1 Old cultures of the iuternal watermelon fungus, when transferred from sterile 

 horse dung to potato broth, produced, along with the small elliptical non-septate 

 eonidia, all sizes and shapes up to those which were luunlate, 3 to 4 septate, and 

 38 M long, hut none 50/1 long could be obtained in this medium, nor were any of 

 them distinctly salmon colored. These cultures were all derived originally from a 

 single non-septate microconidium separated out by the poured-plate method. 



When hothouse watermelons were grown iu soil infected with pure cultures of 

 the internal melon fungus the macroconidial stage frequently appeared on the sur- 

 face after the wilting and death of the ])lants. One of these big, lunulate, 3-septate, 

 external eonidia was separated from its fellows and cultivated for a long time, first 

 in sterile horse dung and subsequently in potato broth, in the same manner as the 

 preceding. The mycelium derived from this spore produced great numbers of micro- 

 conidia, indistinguishable from Ihose borne on the internal mycelium. At least 

 three-fourths of the whole number of spores were of this type, but there were also 

 plenty of typical macroconidia, spores 40 to 50 /i by 4 to 6 /i, lunulate, and 3 to 

 5 septate. Betweeu these two extremes there were all possible gradations, showing 

 clearly that little dependence can be placed on statements respecting shape and size 

 of spores in the ordinary systematic descriptions of Fusaria. Color is likewise 

 misleading, since the same eonidia bed may be white, pinkish, pale salmon, deep 

 salmon, etc., according to its age and the character of the substratum on which it 

 has developed. 



